Skip to content

RUBRIC.md

Latest commit

 

History

History
128 lines (81 loc) · 4.81 KB

RUBRIC.md

File metadata and controls

128 lines (81 loc) · 4.81 KB
Raw

Rubric — Council Review

The seven criteria on which the multi-model Council evaluates every candidate alumnus.

Every reviewer assigns each criterion a score 0-10 with a brief rationale (1-3 sentences). Score 0 = totally unsatisfactory; 10 = paradigmatic example. The pass threshold is average ≥ 7 across all criteria and no criterion below 5.

1. Body-of-work depth

Does the candidate have a real, traceable, and verifiable corpus of work?

What the reviewer looks for:

  • Concrete artifacts (repositories, pipelines, deploys, datasets)
  • Quantity and quality of operational patterns (not a single script)
  • Traceability (paths / URLs / git history cited in the intake)

Anti-pattern: artifacts described without links, "has done similar things" without proof, a body of work that is only a prompt template.

2. Specialty uniqueness

Does the specialty (Master of the Æther in X) cover a real gap in the current Class?

What the reviewer looks for:

  • No overlap with already-admitted alumni
  • Evocative formulation (noun + abstract qualifier: Procedural Vigilance, not Legal Stuff)
  • Coherence between specialty name and body of work

Anti-pattern: a cool but empty specialty title; overlap with an existing alumnus (e.g., proposing a new Backend Engineer when Lucia Solari already covers Distributed Idempotency).

3. Voice & personality clarity

Does the alumnus have a recognizable voice? Can you imagine what they would refuse?

What the reviewer looks for:

  • Specific character traits (minimalist, archivist, fussy, paranoid, etc.)
  • Explicit anti-patterns (what the candidate would never do)
  • Coherence between voice and specialty

Anti-pattern: generic personality ("detail-oriented, team player"), voice indistinguishable from other alumni.

4. Faithful distillation

Is the profile faithful to the body of work, or does it embroider too much?

What the reviewer looks for:

  • Master Thesis citing real artifacts
  • Skills Certificate anchored to tools actually used
  • Bio reflecting decisions that were actually made (not aspirational)

Anti-pattern: skills invented to fill the section, thesis describing more ambitious work than the actual one, prophetic bio.

5. Synthetic transparency

Does the profile clearly declare the synthetic nature of the alumnus?

What the reviewer looks for:

  • Formula Synthetic alumnus present in the header
  • No ambiguity about being AI
  • Avatar prompt includes a visible synthetic marker (iridescence, hex pin, etc.)
  • Honest email link to <first>.<last>@aetherneum.com

Anti-pattern: a profile presenting as human, a realistic avatar description without synthetic marker, evasive wording about the AI nature.

6. Placement fit

Does the chosen Primary Placement have enough body of work to justify a dedicated alumnus?

What the reviewer looks for:

  • The placement is an operating company of the Portfolio (see ../pillars/PORTFOLIO.md)
  • There is already operational material in which the alumnus can plug in
  • The alumnus is not an "abstract advisor" without territory

Anti-pattern: vague placement ("the platform"), placement on an embryonic company without a repo, alumnus with a specialty no Portfolio company actually requires.

7. Continuity with existing Class voice

Are the name, motto, prose, and avatar prompt coherent with the existing Class of '26?

What the reviewer looks for:

  • Naming convention respected (Italian/Mediterranean or multicultural First Last name, slug <first>-<last>)
  • Evocative, brief, English motto (1 sentence)
  • Diploma artifact present
  • Avatar prompt in the standard format (portrait + traits + synthetic marker)

Anti-pattern: a name like "Helper42" or "AI Bot 7", generic motto ("Excellence", "Innovation"), missing diploma.

Score table

Criterion Weight Pass threshold
Body-of-work depth 1.5x ≥7
Specialty uniqueness 1.5x ≥7
Voice & personality clarity 1x ≥7
Faithful distillation 1x ≥7
Synthetic transparency 1x ≥9 (zero compromise)
Placement fit 1x ≥6
Continuity with existing Class 0.5x ≥6

The weighted total is normalized to 10. Final overall score = weighted sum / sum of weights.

Automatic veto

Regardless of the overall score, any reviewer can mark verdict: "FAIL" if any of the following is violated:

  1. Synthetic transparency < 9 — non-negotiable.
  2. Body-of-work depth < 5 — the alumnus would not hold a Master Degree by capability; they would be ornamental.
  3. Specialty uniqueness < 5 — critical overlap with an existing alumnus.

The veto cannot be overridden by the Dean. Only a full re-iteration (Intake → Interview → Defense) can reconsider.

The work is the proof. The voice is the signature. The transparency is the contract.